Why contracts emit events in the first place
Indexers need something reliable to watch for, and raw transaction data alone doesn't cleanly provide that. A transaction record shows that a call was made to a contract with certain inputs, but it doesn't necessarily spell out, in an easily interpretable way, what actually happened as a result. Did a token transfer occur? Did a loan get liquidated? Was a vote cast? Inferring that purely from low-level execution data would be fragile and expensive to redo for every contract and every kind of action. Smart contract events, sometimes called logs, exist to solve exactly this problem: contracts can explicitly emit an event during execution to record that something meaningful just happened, in a structured, predictable format.
Emitting an event is deliberately cheap and simple compared to other kinds of on-chain state changes, which is part of why contract developers use events so liberally. A token contract emits an event on every transfer, a lending contract emits one on every deposit, borrow, or liquidation, a governance contract emits one on every vote cast. None of that is incidental, it's a structured record left specifically so that outside systems, indexers most of all, don't have to guess what happened, they can just read it directly off the event.